Filed in
rails, ruby |
13 March, 2008
Thin is getting some attention, so I thought we would give it a try.
Installation is just a matter of gem install thin.
Run it with something like
thin -e production -s 6
That’s 6 servers running on 0.0.0.0:3000 to 0.0.0.0:3005
Look at the examples if you need to make a monit recipe.
One thing we have is some code to make individual log files for each server instance. This is how it was with mongrel – we put this code in environment.rb inside the Rails::Initializer block:
if ENV['RAILS_ENV'] == 'production'
if defined?(Mongrel::HttpServer)
ObjectSpace.each_object(Mongrel::HttpServer) {|i| @port = i.port}
@port = "unknown" unless @port && @port.to_i > 0
config.logger = Logger.new(File.expand_path(
RAILS_ROOT+"/log/#{ENV['RAILS_ENV']}.#{@port}.log"), 2, 25000000)
end
end
Somthing very similar will work with thin:
if ENV['RAILS_ENV'] == 'production'
if defined?(Thin::Server)
ObjectSpace.each_object(Thin::Server) {|i| @port = i.backend.port}
@port = "unknown" unless @port && @port.to_i > 0
config.logger = Logger.new(File.expand_path(
RAILS_ROOT+"/log/#{ENV['RAILS_ENV']}.#{@port}.log"), 2, 25000000)
end
end
With this code in place you will get individual log files named production.3000.log, production.3001.log etc.
Finally we were seeing these errors:
terminate called after throwing an instance of 'std::runtime_error'
what(): unable to delete epoll event: Bad file descriptor
This is a known problem not with thin, but with EventMachine. Grab an updated gem like this:
gem install eventmachine --source http://code.macournoyer.com
Filed in
rails, ruby |
25 February, 2008
Sometimes you just don’t want to instantiate a bunch of ActiveRecord objects for getting some simple information from the database. You might save on memory and it’ll be faster.
I made a very simple plugin that enables you to get all the values for one column in a table with a simple class method named according to the column – so for instance you can just say People.first_names and you will get all of them in a hash – key is the id, and value is the first_name.
View it here – and install with:
script/plugin install http://pennysmalls.com/rails_plugins/column_as_array
Here’s the README:
ColumnAsArray
=============
This extension allows you to get all the values for a particular column from a table in one hash with a simple call using the column’s name in plural. The resulting hash keys are the ids, and the values are the column values.
The idea is that in cases were only one column of data is needed from a table, but all or nearly all items in the table need to be read – perhaps for listing purposes – then we don’t need to instantiate a whole bunch of ActiveRecord objects just for this. This can save a great deal of heap space if there are many columns in a table.
Example
=======
>> Widget.names
=> {7135=>"Big widget", 33=>"Old widget", 100=>"Fast widget"
>> Thing.updates_ats
=> {2865=>"2008-02-04 09:57:55", 2344=>"2008-01-31 10:24:31", 1823=>nil, 260=>nil}
We use singularize to get the column name from the method name, and this works so that column names that are already plural need to have an extra ’s’:
>> Client.format_choicess
=> {7135=>"csv,xml", 33=>"xml", 100=>"csv,xls", 2110=>"csv", 167=>"xml"}
Filed in
rails, ruby |
28 January, 2008
Sometimes ActiveRecord wraps your data up too much, and you don’t want or need all that convenient but processor-cycle consuming abstraction.
We had a case where we wanted to delete nearly 3 million records from our database. The conditions for deletion were a little complex, and writing a single SQL query for it was not practical.
So it’s nice to do it in Ruby, but instantiating 3m ActiveRecord objects is just not an option. Well, we didn’t need to – you can get data more directly from the database by talking directly to the (mysql in our case) connection.
Here’s how to get a list of ids from the database that were last updated last year.
result = ModelClass.connection.execute("SELECT id FROM table
WHERE updated_at < '2008-01-01'")
ids = []
result.each {|row| ids << row[0].to_i}
Then you can go through the ids and do what you like with them, perhaps compare them to other lists of ids from other tables (which is what we did).
Deleting the records is easy, just use ModelClass.delete(the_id) which works without instantiating an object. Use with care!
Filed in
rails, ruby |
7 January, 2008
Apart from well known stuff like start_form_tag being deprecated, these things broke for us with Rails 2.0.2:
1. Super is no longer called in tests. Use setup_with_fixtures instead, which will work in rails 2.0.2 and future versions where the bug is fixed.
2. The paths to partials used by ActionMailer have changed – now we must use “controllername/partialname” as opposed to “../controllername/partialname”
3. The handling of plusses in urls has changed. It probably changed in this changeset. The result is that passing a + in a url path no longer gets translated to a space in the handling done by rails. So now we have correct handling according to RFC2396.
Filed in
rails, ruby |
26 December, 2007
We have a nice new server which runs 64 bit fedora linux. It’s quick and good. But alas FreeImage won’t compile on this architecture. So bang goes any chance of using Image science.
But really I am only interested in finding the dimensions of my images. And it shouldn’t be hard. I found this post about using the gd library to resize images. Well I don’t need to resize images, but I thought it was a good approach, so I hacked it to do what I wanted.
Here is my code (credit to Damien Tanner for the original code). Very short and simple, and you just need the gd library installed together with the header files. On my system ‘yum install gd-devel’ did the trick.
if !ENV['HOME']
ENV['INLINEDIR'] = RAILS_ROOT + "/tmp"
end
require 'inline'
class ImageInfo
SUPPORTED_FORMATS = %w(jpg jpeg png gif)
def initialize(filename, type=nil)
@filename = filename
@type = SUPPORTED_FORMATS.index(type || @filename[/[^\.]*$/].downcase)
end
def height
unless @height
image_size
end
@height
end
def width
unless @width
image_size
end
@width
end
def fetch_image_size(filename, image_type); end
def image_size
if @type
fetch_image_size(@filename, @type)
else
raise "Unknown type of image"
end
end
inline do |builder|
builder.include '"gd.h"'
builder.add_link_flags "-lgd"
builder.c <<-"END"
void fetch_image_size(char *filename, int image_type) {
gdImagePtr im_in;
FILE *in;
in = fopen(filename, "rb");
/* Support diff image types: jpg jpeg png gif */
switch(image_type) {
case 0:
case 1: im_in = gdImageCreateFromJpeg(in);
break;
case 2: im_in = gdImageCreateFromPng(in);
break;
case 3: im_in = gdImageCreateFromGif(in);
break;
}
fclose(in);
if (im_in) {
rb_iv_set(self, "@width", INT2FIX(im_in->sx));
rb_iv_set(self, "@height", INT2FIX(im_in->sy));
}
}
END
end
end
To use it, do something like this:
begin
iinfo = ImageInfo.new(filepath)
my_image_width = iinfo.width
my_image_height = iinfo.height
rescue Exception=>e
# check errors
end
Filed in
rails, ruby |
4 December, 2007
Image Science is a small plugin that does just the stuff you need from RMagick, without all the weight that RMagick brings with it. It uses FreeImage to do the manipulations. (By the way, the MacPorts installation of that completely failed on my Leopard Mac, I advise you compile it by hand.)
I was getting these kind of errors from my dev site:
SystemExit (exit):
/usr/local/lib/ruby/gems/1.8/gems/RubyInline-3.6.5/lib/inline.rb:70:in `exit'
/usr/local/lib/ruby/gems/1.8/gems/RubyInline-3.6.5/lib/inline.rb:70:in `rootdir'
/usr/local/lib/ruby/gems/1.8/gems/RubyInline-3.6.5/lib/inline.rb:84:in `directory'
/usr/local/lib/ruby/gems/1.8/gems/RubyInline-3.6.5/lib/inline.rb:258:in `so_name'
/usr/local/lib/ruby/gems/1.8/gems/RubyInline-3.6.5/lib/inline.rb:294:in `load_cache'
/usr/local/lib/ruby/gems/1.8/gems/RubyInline-3.6.5/lib/inline.rb:678:in `inline'
/usr/local/lib/ruby/gems/1.8/gems/image_science-1.1.3/lib/image_science.rb:84
Not nice. Curiously everything worked fine on the production system. A bit of digging in the code was required. Image Science uses a gem called RubyInline to allow it to compile and install C extensions to Ruby on the fly. And RubyInline places these extensions in a directory called “.ruby_inline”. It chooses this directory from either the environment setting INLINEDIR, or if that isn’t set then HOME.
And there is the problem – my dev system was running as apache which does not have a home directory.
I added this kind of code:
if RAILS_ENV == "development"
ENV['INLINEDIR'] = RAILS_ROOT + "/tmp"
end
And I made sure that tmp in my rails dir was writeable by apache (but do not make it world writable, that won’t work either).
Filed in
IOCCC |
7 November, 2007
Here is my video of the IOCCC presentation at the Vintage Computer Festival at the Computer History Museum. Low quality I’m afraid – just recorded with the built in mic and camera on my Mac.
Filed in
OSX, rails |
28 October, 2007
Some small breakage in Leopard – it seems a couple of parts of X11 have been removed. Starting my rails setup, which has GraphicsMagick installed (read here for how), results in this:
dyld: NSLinkModule() error
dyld: Library not loaded: /usr/X11R6/lib/libdpstk.1.dylib
Referenced from: /usr/local/lib/ruby/gems/1.8/gems/rmagick-1.15.9/lib/RMagick.bundle
Reason: image not found
Trace/BPT trap
Those dylib files do not exist any more. But… I have a backup (I hope you do). I ended up doing this:
$ cd /Volumes/BackupDisk/usr/X11R6/lib
$ sudo cp libdps*1.0* /usr/X11R6/lib/
$ sudo ln -s /usr/X11R6/lib/libdpstk.1.0.dylib /usr/X11R6/lib/libdpstk.1.dylib
$ sudo ln -s /usr/X11R6/lib/libdps.1.0.dylib /usr/X11R6/lib/libdps.1.dylib
All fixed.
Without a backup you are going to need to find those two files from somewhere. I’m not sure I can legally post them here, but email me if you are stuck.
Filed in
rails, ruby |
25 October, 2007
Not only does exception notifier not work, you probably don’t know that it doesn’t work either. All your code has just become super-exception-free all of a sudden.
Ah, wishful thinking.
If you have installed Ruby 1.8.6p111 then you will want to take note of the comments here.
Just modify the file vendor/plugins/exception_notification/views/_environment.rhtml to say:
* <%= "%-*s: %s" % [max.length, key, @request.env[key].to_s.strip] %>
instead of
* <%= "%*-s: %s" % [max.length, key, @request.env[key].to_s.strip] %>
Filed in
rails |
19 October, 2007
We have a problem – we really do. Each time a user requests a page they have to make 50 http requests just to get back a “Not modified” message from the web server. Their browser is asking about every little image and css file and so on. Every one on the page. Those messages are small, but they add up. And as we know, “make fewer HTTP requests” is Steve Souder’s number one rule for speeding up your website.
What we need to do is to add an Expires header to our static content – one that gives a time a long way in the future so that the browser knows that the content should stays in the cache -and not be fetched again, or even checked (if it has been updated) from the server. But what if we want to change a css file or a js file? The user will get the old one from their own cache. No good.
Well Rails has a mechanism to prevent this – it adds a 10 digit number on the end of each url (as generated by image_tag, stylesheet_link_tag, or javascript_include_tag for instance) in the query string. That number is based on the file modification time – so when the file is updated then the URL will change. It’s like having a remote way to expire the item in the browser’s cache.
So it’s simple right? Just turn on mod_expires in Apache? Not so fast. What about those images and items that do not have the query string? We don’t want to send an expires header for those. Definitely not. (And you will likely have some – for instance images that are referenced from your css files.) If you do, the user is stuck with their cached version even if you change the file on the server.
So we need some way of selectively turning on the expires header in apache.
One way [Danny Burkes] (look at the update at the bottom) is to segregate by directory what you want to expire and what not. But this seems a bit clumsy to manage. (Also, side issue, I am not totally convinced that the munging of the urls provided by the plugin is needed – at least section 13.9 of the HTTP 1.1 spec seems to suggest that content with query strings will be cached fine if an explicit expires header is given.)
Actually you can do what you need with one symbolic link and some apache magic. The obvious magic won’t work – you can’t detect what’s in a query string using a LocationMatch or FilesMatch container. But we can get around this with a rewrite rule, and a directory container.
This is from my apache httpd.conf file:
# add something we can do a directory match on
RewriteCond %{QUERY_STRING} ^[0-9]{10}$
RewriteRule ^(.*)$ /add_expires_header%{REQUEST_URI} [QSA]
# the add_expires_header directory is just a symlink to public
<Directory "/path/to/rails_app/public/add_expires_header">
ExpiresActive On
ExpiresDefault "access plus 10 years"
</Directory>
This detects those query strings (we assume you don’t use 10 digit query strings for anything else), and adds a directory on the front of the path.
We use this as something we can detect in a Directory container. And in there we turn on the expires header.
We need one more thing:
cd /path/to/rails_app/public
ln -s . add_expires_header
The symbolic link doesn’t go anywhere, and that’s just what we want. All the images and css files and whatnot will be found in their usual places. It’s pretty unobtrusive -you don’t need to change anything in your app to start to benefit from the expires header.
It’s a big benefit – we have literally gone from 50 HTTP requests per page to about 16. And with some tweaking we’ll get it down more – some of those are from references to images in css, but a few are due to us not using urls generated by rails for static content. And we can fix those.
« Previous Page — Next Page »
RSS